The Raven's 2 Clinical Edition developed by Pearson features an item bank comprised of questions numbering in the hundreds.
Because of this each test differs from the other with possible combinations of different questions numbering in the millions. This was developed for test security in as well as practice effect mitigation. A friend of mine has taken this test a multitude of times resulting in variable scores with a large spread. He knows he's not supposed to.
The scores of a three month long period are as follows.
144, 133, 130, 150, 140, 140, 141, 130, 135
In that order
Why is there so much variability in scores? Which score most accurately reflects this persons intelligence quotient? Is this correctly measured within the confidence interval or margin of error of Raven's 2.